How to Shuffle a DataFrame in pandas: Simple Guide

To shuffle rows in a pandas DataFrame, use the sample() method with frac=1 to return all rows in random order. For example, df.sample(frac=1) returns a shuffled DataFrame.

Syntax

The main method to shuffle a pandas DataFrame is sample(). It has these key parts:

• frac=1: means return 100% of rows in random order.
• random_state: sets a seed number to get the same shuffle every time (optional).
• replace=False: ensures rows are not repeated (default behavior).

python

df.sample(frac=1, random_state=None, replace=False)

Example

This example shows how to shuffle all rows of a DataFrame and reset the index for a clean result.

python

import pandas as pd

data = {'Name': ['Alice', 'Bob', 'Charlie', 'David', 'Eva'],
        'Age': [25, 30, 35, 40, 45]}
df = pd.DataFrame(data)

# Shuffle all rows
shuffled_df = df.sample(frac=1, random_state=42).reset_index(drop=True)
print(shuffled_df)

Output

      Name  Age
0    David   40
1      Eva   45
2    Alice   25
3  Charlie   35
4      Bob   30

Common Pitfalls

Some common mistakes when shuffling a DataFrame:

• Not using frac=1 will return only a fraction of rows, not a full shuffle.
• Forgetting to reset the index after shuffling keeps the old row numbers, which can be confusing.
• Not setting random_state if you want reproducible shuffles.

python

import pandas as pd

data = {'Name': ['Alice', 'Bob', 'Charlie'], 'Age': [25, 30, 35]}
df = pd.DataFrame(data)

# Wrong: returns only 50% rows shuffled
partial_shuffle = df.sample(frac=0.5)
print(partial_shuffle)

# Right: shuffle all rows
full_shuffle = df.sample(frac=1).reset_index(drop=True)
print(full_shuffle)

Output

    Name  Age
1    Bob   30

      Name  Age
0  Charlie   35
1    Alice   25
2      Bob   30

Quick Reference

Parameter	Description	Default
frac	Fraction of rows to return in random order	1 (all rows)
random_state	Seed for reproducible shuffling	None
replace	Allow sampling with replacement	False
reset_index(drop=True)	Reset index after shuffle to clean row numbers	Must call explicitly

Key Takeaways

Use df.sample(frac=1) to shuffle all rows of a DataFrame.

Set random_state for reproducible shuffles.

Call reset_index(drop=True) after shuffling to reset row numbers.

Avoid using frac less than 1 if you want to shuffle the entire DataFrame.

Sampling with replace=True can duplicate rows, usually not desired for shuffling.